An experiment is described where six on-line techniques for monitoring milk coagulation (NIR diffuse reflectance, NIR transmission (two wavelengths), hot-wire, torsional vibration and tuning-fork vibration) were compared with rheological measurements of curd firming (elastic shear modulus) under a range of coagulation rates, effected by varying enzyme level in fresh whole milk. Viscous and elastic moduli were measured during renneting and were analysed to determine their usefulness for determination of a gel point. G' and G" each gave a suitable measure of gel time as indicated by compliance with Holter's Law. The correlation of gel points determined by G" with those determined by G' showed that a sensor which detects changes in viscosity per se works in principle as a gel point detector in cheesemilk. Gel times determined from GN were slightly longer than those determined from G' which in turn were slightly longer than those determined from tan delta, with; the differences becoming larger at lower enzyme levels, i.e. longer gel times. Gel times based on G" were just as good predictors of cutting time as those based on G'. In addition the asymptotic behaviour of tan delta showed that a viscosity sensor can determine an optimum point of curd-cutting, which is usually defined in terms of a large value of G'. All on-line techniques modelled reological gel time (by rheometer) with an accuracy of ca. 1 min. and predicted curd cutting time with an accuracy of ca. 2 min.